In a typical voice coil actuator, an electrical current conductive coil is suspended at a zero current bias position within a magnetic field formed in a gap of a permanent magnet core. It is highly desirous that the flux path of the field within this gap be optimally radial with respect to the axis of the coil so that when an externally applied current conducts through the coil, the coil will be displaced axially from its zero current bias position in an amount linearly proportional to such current. However, as will be seen from the following discussion, the flux path is, in the known prior art voice coil actuator, not at an ideal radial orientation with respect to the coil.
Referring to prior art FIG. 1, there is shown a long coil actuator 10 which is one particular type of prior art voice coil actuator. The construction of its core 12 is based on an axially polarized cylindrical magnet 14 which has a first end face 16 at a first magnetic polarity and a second end face 18 at a second, opposite magnetic polarity. The first end face 16 of the magnet 14 is coaxially mounted on a disk shaped base plate 20 of the core 12. The base plate 20 is formed from magnetic flux conductive material and has a diameter commensurate with the outer diameter of the magnet 14. Mounted on the base plate 20 coaxially within the magnet 14 is a rod 22 of magnetic flux conductive material wherein the rod 22 has a diameter less than the inner diameter of the magnet 14 so that a space therebetween remains. The distal free end 24 of the rod 22 is elevationally commensurate with the second end face 18 of the magnet 14. Completing the core 12 construction, a ring shaped first pole piece 26 and a disk shaped second pole piece 28 are coaxially mounted respectively to the second end face 18 of the magnet 14 and the distal end 24 of the rod 22. Each pole piece 26,28 is formed from magnet flux conductive material. The first pole piece 26 has an inner diameter less than the inner diameter of the magnet 14 and the second pole piece 28 has a diameter greater than the diameter of the rod 22 with each of these diameters being selected so that a gap remains between the respective facing sides 30,32 of each pole piece 26,28.
A coil 34 is mounted to the above core 12 construction and coaxially suspended within the gap. The overall length of the coil 34 and the height of the gap are selected so that as the coil 34 is moved through its total stroke, defined as the maximum positive and negative axial deviation from its zero current bias position, the number of coil turns within the gap, defining the effective length, L, of the coil 34 remains constant. In the long coil actuator 10, an equal length of the coil 34 extends axially from each end of the gap when the coil 34 is in the zero current bias position as best seen in prior art FIG. 1. The positive and negative stroke limit is then equal to the axially projecting length.
In any prior art voice coil actuator, this condition on the constancy of the coil turns, or the constancy of the effective length, L, within the gap is necessary to maintain the linearity of the relationship between magnetic force, F, on the coil with respect to current, i, within the coil so that F .varies. i. Since F=iLxB, and the magnetic flux density, B, which is assumed to be uniform and radially confined within the gap, it is seen that the force, F, is linearly dependent on the coil current, i, or F=LBi, wherein LB is the constant of proportionality. Since the current is assumed to be perpendicular to the uniform radially confined flux, the cross product factor of sine becomes equal to one. The above statement that the gap height and the effective coil length, L, are identical for the long coil actuator 10 from which the linear dependency F .varies. 2 is derived may only be made if the magnet flux density, B, is assumed to be radially confined and uniformly distributed within the gap.
Another type of known prior art voice coil actuators is known as a short coil actuator. In the short coil actuator, the axial length of the coil is less than the height of the gap and, at zero current bias is centered therein. The positive and negative stroke limit is then equal to one-half of the gap height less the coil length so that the coil is always confined within the gap. Based on the above assumption, the effective length, L, of the coil is then equal to its actual length.
Considering only the geometry of the construction of the long and short coil actuator designs, it would appear that the above condition for linearity exist as long as the magnetic flux density within the gap is assumed to be radially confined and uniformly distributed therein. However, it will be shown that in the prior art voice coil actuators, this assumption is not true. It will therefore become apparent that for either the long coil or short coil actuator, the force, F, is a function of each of the current, i, length, L, and flux density, B, or F=.function. (i,L,B) wherein .differential.F/.differential.L and .differential.F/.differential.B are nonzero quantities as the coil moves axially due to magnetic flux leakage at the edges of the gap and nonuniformity of flux density along the length of the gap, as set forth in greater detail hereinbelow. This nonzero dependency of the force, F, on each of the effective coil length, L, and flux density, B, causes a nonlinear response to the coil current, i, and results in harmonic distortion of the coil actuation. This harmonic distortion is one disadvantage and limitation of the prior art voice coil actuator in that the sound produced by the actuation of the voice coil driving a speaker cone is not a pure analog of the coil current. Another disadvantage and limitation is a reduction of efficiency due to leaking magnetic fields not contributing to the flux passing through the coil.
The assumption on the radial confinement and uniform distribution of the flux density within the gap is false for two reasons which although demonstrated herein below in reference to the long coil actuator 10 are also applicable to the short coil actuator. First, since the respective lower 36, 38 and upper 40, 42 edges of the first and second pole pieces 26,28 present a discontinuity, the flux density fringes outside of the gap from the edges 36,38,40,42 of the pole pieces 26,28 such that the flux is not radially confined within the gap. Secondly, since magnetic flux follows the path of least resistance, the flux density in the gap is therefore greatest at the lower edge 36 of the first pole piece 26 where the first pole piece 26 is adjacent the magnet 14 and decreases toward the upper edge 42 of the second pole piece 28 such that the flux is not uniformly distributed. The nonuniformity of distribution also occurs at the fringing fields at either end of the gap.
In either the long or short coil actuator, the nonuniformity of the flux distribution results in the flux which interacts with the coil to be dependent on the instantaneous position, x, of the coil in the gap as determined at the center of the coil. Since the coil has a finite length within the gap, the current through each loop will be effected by a different value of flux density. The total force, F, acting on the coil will therefore be a summation over the length of the coil of the force developed by the current in each loop of the coil interacting with the flux resulting in the expression F=.intg.i(x).sup.xB(x)dx, where the lower and upper limits of integration are x-L/2 and x+L/2, respectively. Since the coil is continuous over its length, the current in each loop is identical, or i(x)=i so that the above expression simplifies to F=i .intg.B(x)dx. Therefore, .differential.F =(.intg.B(x)dx).differential.i+(i.intg.B(x)dx).differential.B.
Therefore, it is seen that the change of force, .differential.F, is not linear with respect to the change of current, .differential.i, but depends on the coil position, x, within the gap, the length of the coil, L, since these terms are within the limits of integration, and the variation of the flux density, B, over the length of the coil. This nonuniformity is especially prevalent in the short coil actuator since the gap height being much greater than the coil length causes the coil to be exposed to a greater degree of nonuniformity of the flux density along the length of the gap. An additional nonlinearity is introduced in the long coil actuator since the coil at either stroke extreme is removed from a fringing field external of the gap, thereby making the effective coil length, L, also a function of the coil position, x, or L=L (x) , which therefore introduces a nonlinear variable into the limits of integration.
In addition to the above described nonlinearities of the known voice coil actuator, a further limitation and disadvantage is the magnetic flux leakage external of the core. For example with reference to prior art FIG. 1, the flux will also fringe from each pole end face 16,18 of the magnet 14 external from the core 12. This fringing flux is usually caused by saturation at the core. First, this external fringe results in a waste of useable flux, thereby reducing conversion efficiency of the prior art voice coil actuator. Secondly, this external fringe can also detrimentally interact with nearby electronic circuitry, which often requires the prior art voice coil actuator to be heavily shielded which adds to its size, weight and cost.